PROJECT SUMMARY Electronic Health Records (EHRs) can improve the quality of healthcare delivery in the United States, by providing automated best-practice reminders to clinicians and patients. However such functionality is currently limited to narrow areas of clinical practice, as existing decision support systems can process only structured data, due to lack of a suitable framework and concerns about accuracy and portability. Preliminary work by the PI has shown that rule-based approach can be used to develop broad-domain reminder systems that can utilize free-text in addition to the structured data. The PI has developed prototype systems for cervical and colorectal cancer prevention. These systems consist of rule-based composite models of national guidelines, and rule-based Natural Language Processing (NLP) parsers. The NLP parsers extract the patient variables required for applying the guidelines. However further research is needed to extend the systems and to ensure their accuracy for clinical deployment. In the mentored phase, the PI will collaborate with clinicians to extend and iteratively optimize and validate the systems, and will make them available in open-source so that they can be adapted for deployment at other institutions (aim 1 - K99). In the independent phase, the PI will research methods to facilitate rapid development, deployment and cross- institutional portability of similar systems. Specifically, the PI will develop a hybrid design for the parsers and investigate domain adaptation and active learning methods, for reducing the manual effort for development and adaptation of the NLP parsers (aim 2 - R00). To enable other researchers to reuse the developed methodologies and software resources, a toolkit will be developed that will support the construction and deployment of similar systems (aim 3 - R00). The toolkit will consist of user-friendly tools and templates to replicate the processes engineered in the case studies, and will build on the SHARPn data normalization tooling and other open-source tools. The independent phase will be in collaboration with Intermountain Healthcare. The PI's career goal is to become a scientific leader in clinical informatics with a focus on optimizing clinical decision making. The PI has strong background in clinical medicine and medical informatics, and will receive mentoring from Drs. Hongfang Liu, Christopher Chute, Robert Greenes and Rajeev Chaudhry, who have complimentary areas of expertise. The mentored (K99) phase will be for 2 years at Mayo Clinic Rochester, wherein the PI will undertake courses on decision support and will get mentored training in NLP and health information standards. This will prepare the PI for independent research in R00 phase on portability and tooling. Completion of the proposed work will enable the PI to seek further funding for piloting clinical deployment of the developed systems, measuring their clinical impact, and for scaling the approach to other clinical domains and institutions. The career grant will enable the PI to establish himself as an independent investigator and to make significant contributions towards advancing clinical decision support for improving care delivery.